A configuration error inside Microsoft’s Azure cloud knocked large parts of the British internet offline for several hours on Wednesday, leaving travellers unable to load Heathrow’s booking pages, bank customers staring at blank NatWest screens, and supermarket shoppers unable to reach Asda’s grocery site. Microsoft said the disruption stemmed from “DNS issues” on Azure, the same class of failure that took down Amazon Web Services last week.
The company’s status page first flagged degraded service around 16:00 GMT. Downdetector logged thousands of outage reports in the United Kingdom, the United States, and parts of Europe as domain name lookups failed or timed out. Microsoft 365 users also reported delays sending mail through Outlook. By 21:00 GMT the firm said it had rolled back the offending change and that “many websites that went down were once again accessible.”
Which UK services were hit
Heathrow Airport confirmed its public website was unavailable during the peak of the fault, though flight operations continued and passengers were directed to airline apps and on-site staff. NatWest Group said online banking and parts of its mobile service were intermittently unreachable, while Marks & Spencer and mobile network O2 reported similar symptoms. The BBC’s own technology desk noted that Minecraft, which runs on Azure, was among the globally visible casualties.
Because Azure underpins so many third-party sites indirectly, the outage rippled beyond brands that advertise Microsoft contracts. Retailers, insurers, and local councils that host customer portals on Azure-backed infrastructure saw error pages even when their own data centres were healthy. Cyber security agencies did not classify the event as a malicious attack; Microsoft attributed it to an internal software update that propagated bad DNS records.
Why DNS failures cascade
Domain name system servers translate human-readable web addresses into numeric routes. When a major provider mis-publishes records, browsers cannot find the correct server even if the application behind it is running. That makes DNS a single point of failure for cloud-heavy economies such as Britain’s, where government digital services and the NHS increasingly rely on hyperscale vendors.
Regulators have warned about concentration risk since a 2024 incident at a smaller provider briefly disrupted GP appointment systems. Wednesday’s episode was wider but shorter, ending once Microsoft engineers reverted to a known-good configuration. The UK’s National Cyber Security Centre said it was monitoring the situation and stood ready to support critical sectors, though it had not activated emergency protocols.
What businesses and passengers should expect
Airlines advised Heathrow passengers to arrive with printed boarding passes or airline-app confirmations in case web check-in remained flaky. Banks told customers to use telephone banking if apps failed, and NatWest said no evidence suggested account data had been exposed. Microsoft promised a post-incident review but did not say when a full root-cause report would be published.
For corporate technology chiefs, the outage is another reminder to maintain offline fallbacks and multi-cloud DNS redundancy. For the public, it was a frustrating afternoon of refresh buttons — a localized breaking story with global roots in a single US software rollout.
